relative spatial observations

As this seems to be the trend for inputs in many of the ML papers we're seeing these days (so-called "vector representations" of the spatial scene from an egocentric perspective), I propose we add the following to our standard Observations:

To the VehicleObservation and EgoVehicleObservation classes, I propose we add:

lane_relative_heading - which is the result of applying min_angles_difference_signed() to the vehicle's and lane's local headings. (i.e., it's the difference between the vehicle's current heading and the way it's supposed to be going at this point in the lane, if it were centered in the lane.)

Then, in the Observation class, I propose we change the type for neighborhood_vehicle_states to be Optional[Dict[str, Tuple[VehicleObservation, RelativeObservation]]]. The Dict here would be keyed on vehicle_id. The new type RelativeObservation could look like:

@dataclass
class RelativeObservation:
    """Information about the relationship between the vehicle making observations and a target object."""

    target_id: str
    """Unique id of the target object."""
    range: float
    """The 'as-the-crow-flies' Euclidian distance to the target object."""
    bearing: float
    """The angle to the target object relative to north."""
    relative_bearing: float
    """The angle to the target object relative to my own heading."""

Note that relative_bearing is redundant and may not be needed (as it can be computed from bearing and heading) but this seems to be the most useful info here, so we might as well compute it. (... especially if we also consider keeping track of, and reporting, the rate-of-change of the relative_bearing over a time window, which is something very useful to avoiding collisions!)

We may also consider adding other fields to RelativeObservation including relative_speed relative_acceleration (or vectorized versions of these) and even estimated_time_to_contact (although computing this one requires making several assumptions that might be hard to justify/explain).

If we want flexibility (for example, if the overhead is too high), the NeighborhoodVehicles class in agent_interface.py could be augmented to add a boolean flag include_relative_info which, when True, would include the RelativeObservation objects in the output of the sensor.

huawei-noah / SMARTS

relative spatial observations #1493